摘要
讨论了油气层模式识别的流程和数据处理方法.以白音查干的达尔其为研究工区,T26层位为处理目标,达7井、达8井、达9井、达11井、达12井、达13井为样本井,对孔隙度、含油饱和度等物性数据进行神经网络学习和物性预测.结果表明,神经网络模型对动态变化性强的数据自适应性强,具备外推内插功能,对井间的油气物性变化的预测准确,处理结果合理.
Procedures and processing method of pattern recognition for oil and gas reservoir is discussed. By taking Da'erqi area in Baiyinchagan as a work area of study. T26 horizon as target of processing, Wells Da7, Da8, Da9, Da11, Da12 and Da13 as model wells, neural network is studied and physical properties of physical data including porosity, oil saturation etc are predicted. The result shows that the neural network model has strong self-adaptability to the data with strong dynamic changes, it has the function of interpolation and extrapolation, with precised prediction of interwell physical property change, reasonal processing result.
出处
《江汉石油学院学报》
EI
CSCD
北大核心
2003年第2期45-46,共2页
Journal of Jianghan Petroleum Institute
基金
中国石油化工股份有限公司攻关项目(P01108)
关键词
油气层
模式识别
神经网络
油层物性
油气资源预测
oil and gas layer
pattern recognition
neural network
physical property of reservoir
prediction of oil and gas resource
